The following fails with the V8 bindings because we are not keeping the wrapper for the HTMLFormCollection alive var form= .... form.elements.custom = 42; gc(); shouldBe('form.elements.custom', '42'); This causes a test as part of bug 90194 to fail in V8.
Created attachment 150221 [details] Patch
Comment on attachment 150221 [details] Patch View in context: https://bugs.webkit.org/attachment.cgi?id=150221&action=review > Source/WebCore/bindings/scripts/IDLAttributes.txt:55 > -GenerateIsReachable=|Impl|ImplContext|ImplDocument|ImplElementRoot|ImplFrame > +GenerateIsReachable=|Impl|ImplContext|ImplDocument|ImplElementRoot|ImplFrame|ImplBaseRoot Would you be willing to update the documentation about this property? I don't really understand what these various values mean: https://trac.webkit.org/wiki/WebKitIDL#JSGenerateToJSObject
(In reply to comment #2) > Would you be willing to update the documentation about this property? Sure thing.
Comment on attachment 150221 [details] Patch This LGTM, with the caveat that I don't know whether ImplBaseRoot is a sensible name, but it seems in line with the names of the other values this property can take.
(In reply to comment #4) > (From update of attachment 150221 [details]) > This LGTM, with the caveat that I don't know whether ImplBaseRoot is a sensible name, but it seems in line with the names of the other values this property can take. The names comes from ...::root(impl()->base()). For V8 we don't do the root part but other than that it is the same. I'll try to explain that in the wiki. The values for these make almost no sense without actually looking at the code gen and/or the .h files.
I filed bug 90317 to look into using WebCore::root for V8 too.
Created attachment 150272 [details] Patch for landing
Comment on attachment 150272 [details] Patch for landing Clearing flags on attachment: 150272 Committed r121615: <http://trac.webkit.org/changeset/121615>
All reviewed patches have been landed. Closing bug.